Artificial Intelligence Economic Crimes: Threats and Solutions

Abstract

AI research and regulation seek to balance the benefits of innovation against potential harm and disruption. However, one unintended consequence of recent advancements in AI research is the potential redirection of AI technologies to facilitate criminal activities, which we term "AI crime." In theory, this is made possible by published experiments on automating fraud targeting social media users, as well as demonstrations of AI-driven manipulation of simulated markets. Nevertheless, because AI crime remains a relatively young and inherently interdisciplinary field, encompassing socio-legal studies to standalone formal knowledge, it is challenging to predict the future of AI crime with certainty. This paper provides the first systematic and interdisciplinary analysis of foreseeable AI crime threats, offering law enforcement and policymakers a mix of current challenges and potential solutions.